yh1008 speech-to-text issues - Githubissues

yh1008 / speech-to-text

mixlingual speech recognition system; hybrid (GMM+NNet) model; Kaldi + Keras

http://llcao.net/cu-deeplearning17/project.html

70 stars 19 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

ERROR: phone symbol tables data/lang/phones.txt and exp/mono_ali/phones.txt are not compatible.

#15 yh1008 opened 7 years ago
1
local/score.sh does not exist

#14 yh1008 closed 7 years ago
1
WARNING: triphone decode

#13 yh1008 opened 7 years ago
0
WARNING: arpa2fst

#12 yh1008 opened 7 years ago
2
WARNING: tree has pdf with no stats

#11 yh1008 opened 7 years ago
1
mkgraph.sh: expected data/lang/G.fst to exist

#10 yh1008 closed 7 years ago
2
fail to create mfcc feature

#9 yh1008 closed 7 years ago
1
test data fix_data_dir.sh: no utterances remained: not proceeding further.

#8 yh1008 closed 7 years ago
1
Mal-formed spk2gender file

#7 yh1008 closed 7 years ago
2
utt2spk not sorted

#6 yh1008 closed 7 years ago
2
need to trim audio files?!

#5 yh1008 closed 7 years ago
2
update: MFCC in detail & Kaldi document in cn

#4 wendywangwwt closed 7 years ago
0
How to build language detection model? what would be the unit of an audio?

#3 yh1008 opened 7 years ago
11
How to use Kaldi Recipe

#2 yh1008 closed 7 years ago
0
dataset?

#1 yh1008 closed 7 years ago
3